Acoustical measurements of masks and the effects on the speech intelligibility in university classrooms

Acoustical measurements of three different masks, surgical, KF94, and N95 respirator (3 M 9210+) were performed and compared with the results obtained with no mask on a dummy head mouth simulator to understand the acoustical effects of the three different masks on speech sounds. The speech intelligibility and perceived difficulty of understanding speech sounds with and without an N95 mask were also measured using speech signals convolved with previously measured impulse responses in 12 occupied university classrooms. The acoustic attenuations with the three masks were greatest in front of the talker. The surgical, and KF94 masks resulted in 6–12 dB reductions of high frequency sounds between 2 kHz and 5 kHz, and the N95 respirator decreased sound levels by an additional 2–6 dB at these frequencies. Both surgical, and KF94 masks performed acoustically better at high frequencies between 2 kHz and 5 kHz than N95 mask did. The mean trends of the speech intelligibility test results indicate that young adult listeners at university achieve a mean score of 90% correct at a signal-to-noise ratio (SNR) value of + 8 dBA or higher for no mask conditions, which is a 4 dBA lower SNR value than for N95 mask conditions. The intelligibility scores obtained with N95 mask conditions decreased the correct scores by a maximum of 10% at a SNR of 5 dBA or lower compared to the results obtained with no mask conditions. The perceived difficulty ratings obtained in N95 mask conditions increased the ratings by a maximum of 10% at lower SNR values compared to the results obtained in no mask conditions. Achieving higher SI scores of 95% or more doesn't indicate that the listeners experience no difficulty at all in understanding speech sounds. Higher SNR values are beneficial for achieving better speech communication for both no mask and an N95 mask on a talker in classrooms. © 2021 Elsevier Ltd